In wire communication and wireless communication system, digital modulation is performed in order to transmit information effectively. The one of the most representative digital modulation is Quadrature Amplitude Modulation (QAM). In QAM, digital signal comprising bit row is modulated by changing amplitude and phase of carrier wave.
FIG. 1 illustrates a conventional constellation. In FIG. 1, the constellation having 16 constellation points in conventional 16-QAM is illustrated.
Herein, the constellation is also referred to as I/Q  plot signal space figure. In 16-QAM, 4 bit binary digital signal (modulation symbol) per one constellation point can be transmitted through 16 signal spaces of which the phases and amplitudes are different from one another.
In more detail, bit stream is divided into groups having 1 bit or more than 1 bit, and mapping the each group with one of constellation points is performed, the each of constellation points corresponding to one modulation symbol.
On the other hand, as shown in FIG. 1, as each of constellation points of conventional constellation is arranged to be angular point of square, it is referred to as square constellation (Square QAM).
In square form constellation, distance between two adjacent constellation points is always 2d, and average power consumed by all 16 constellation point is calculated by following equation 1.Pav1=10d2  [equation 1]
However, the conventional art is not optimized in view of power efficiency and modulation method of which the power efficiency is improved is required.
On the other hand, if average power is consumed by equation 1, the average power can be reduced by shortening distance between the constellation points. However, if distance between the constellation points is shortened, modulation symbol error rate becomes higher. Therefore, method that can reduce average power consumption while maintaining modulation symbol error rate is required.